| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Service
d'Hépatogastroentérologie,*
INSERM U
402
and Service
d'Anatomopathologie,
CHU Saint-Antoine,
Paris INSERM U 397,§
CHU Rangueil, Toulouse,
France; and the Department of Radiation Oncology,¶
University of North Carolina at Chapel Hill, Chapel Hill,
North Carolina
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Vascular endothelial growth factor (VEGF) and fibroblast growth factor-2 (FGF-2) are the most potent angiogenic factors identified thus far. Their role in vascular proliferation associated with tumor growth or wound healing has been widely documented in different organs.4 In addition, it has been demonstrated that hypoxia was the main inducer of VEGF expression5 which, in turn, stimulates local proliferation of capillaries to increase oxygen delivery. Besides beneficial effects, this vascular response may nevertheless also result in deleterious effects best illustrated by tumor growth and proliferative retinopathy.6,7
We and others8,9 have recently observed an up-regulation of VEGF in the cirrhotic liver of patients with or without hepatocellular carcinoma, suggesting that this factor might be responsible for cirrhosis-associated angiogenesis. Intrahepatic shunts and capillarization of sinusoids are well established characteristics of cirrhosis that restrict the access of blood solutes to hepatocytes.10-12 In addition, a decrease in the hepatic microvascular perfusion secondary to biliary obstruction in rat has been recently shown to occur long before the onset of cirrhosis, within 7 days following bile duct ligation (BDL).13 Whether these vascular morphological and functional alterations may induce chronic hypoxia in the fibrotic liver and thereby elicit an angiogenic response is unknown.
The aim of the present study was to investigate the role of VEGF and FGF-2 in an experimental model of biliary fibrosis. We next studied the role of hypoxia in triggering VEGF expression and thereby in stimulating angiogenesis in this model.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Male Sprague-Dawley rats were used at a body weight of 200 to 250 g. The common bile duct was ligated as described previously14 while normal and sham-operated rats were used as controls. At various times after surgery (3 days, 1, 2, 3, 4, 5, 6, and 7 weeks), rats were euthanized with an overdose of ketamine (Parke-Davis, Courbevoie, France). Cholestasis was monitored by serum bilirubin level. At the time of sacrifice, arterial and mixed venous blood samples were withdrawn for the measurement of oxygen tension (PO2) and liver tissue was processed as described below.
Histology and Immunohistochemistry
Liver samples were fixed in 10% buffered formalin, paraffin-embedded and sectioned at 4 µm. Tissue sections were stained with hematoxylin-phloxin-safran before standard histology. Immunolabeling was performed using polyclonal antibodies against VEGF (a goat polyclonal IgG: sc-152-G and a rabbit polyclonal IgG: sc-507) (1:100) (Santa Cruz Biotechnology, Santa Cruz, CA) and FGF-2 (1:100) (Santa Cruz Biotechnology) and monoclonal antibodies against von Willebrand factor (vWF) (1:200) (Dako, Glostrup, Denmark). An avidin-biotin-peroxidase technique (Vectastain ABC Kit, Vector, Burlingame, CA) was used for VEGF, FGF-2, and vWF detection. For vWF, immunoperoxidase was performed after microwave antigen retrieval (750 W, 3 x 5 minutes in citrate buffer 0.01 mol/L, pH 6). Before immunostaining, endogenous biotin was blocked using a commercial kit (Eurobio, France) and endogenous peroxidase activity was inhibited in 3% alcoholic hydrogen peroxide for 30 minutes. Color development was achieved with 3-amino-9 ethyl carbazole. The controls were obtained by omitting the first antibody and were all negative. Two independent sections of each sample were evaluated. The cells exhibiting a moderate to intense signal for VEGF and FGF-2 were considered as positive and counted. At least 400 hepatocytes were analyzed in two independent fields. Cell proliferation was assessed by means of a three-step immunoperoxidase method with a monoclonal antibody raised against Ki67 (Novocastra, Newcastle, UK). The cell nuclei were identified as positive or negative and counted. At least 1200 hepatocytes and 600 bile duct epithelial cells were analyzed in three independent fields. Endothelial cell proliferation was assessed by counting the percentage of Ki67 positive endothelial cell nuclei in the periportal vessels of five independent high-magnification (x400) fields per animal. The vascular density in the periportal fibrosis was assessed by determining the count of vWF-labeled vessel sections in 10 successive high-magnification (x400) fields per animal using an eyepiece with a net micrometer (Carl Zeiss, Jena, Germany). The liver samples of at least two rats at each time point (normal rats, 1, 3, and 6 weeks after BDL) were examined. The same procedure was applied to determine the count of bile duct sections. For endothelial cell proliferation and vascular density data, mean values for more than two groups (at each time point) were compared by analysis of variance (Kruskall-Wallis test) and in case of significance, Scheffé or Games-Howell tests were used to detect difference between single groups. All results were expressed as mean ± SEM.
Reverse Transcription Polymerase Chain Reaction Analysis of VEGF Transcripts
Total RNAs were extracted by a guanidinium thiocyanate based method using a commercial kit (Trizol, Gibco BRL). Five micrograms of RNAs were reverse transcribed using a commercial kit (Pharmacia Biotech). Samples of cDNA were subjected to VEGF amplification combined with GAPDH co-amplification by polymerase chain reaction (PCR). The VEGF and GAPDH oligonucleotide primers were designed based on published rat cDNA sequences in EMBL database. The VEGF sense primer was 5'-ACCTCCACCATGCCAAGT-3' (position on cDNA: 54–71) and the antisense primer was 5'-TAGTTCCCGAAACCCTGA-3' (position on cDNA: 602–619). The GAPDH sense primer was 5'-CCATGGAGAAGGCTGGGG-3' (position on cDNA: 335–352) and the antisense primer, 5'-CAAAGTTGTCATGGATGACC-3' (position on cDNA: 510–529). PCR was performed in a 50-µl reaction mixture containing: 10 mmol/L Tris-HCl, pH 9.0); 50 mmol/L KCl; 1.5 mmol/L MgCl2; 0.2 mmol/L dNTPs; 25 pmol/L of each VEGF primer; and 2.5 pmol/L of each GAPDH primer. The PCR conditions were as follow: 94°C for 7 minutes, then 25, 28, 31, 34, and 37 cycles of 1 minute denaturation at 94°C, 1 minute annealing at 57°C, 1 minute 30 seconds extension at 72°C, and a 10-minute terminal extension at 72°C. To monitor the kinetics of PCR product formation, aliquots were withdrawn at different PCR cycles and analyzed by Southern blot after hybridization of the membrane with specific VEGF and GAPDH 32P-labeled probes. PCR products were semiquantified by optical density scanning of the blot.
Western Blot Analysis of FGF-2 Expression
Tissue samples were homogenized using an all-glass homogenizer in ice-cold lysis buffer (Tris 20 mmol/L, pH 7.5; NaCl, 150 mmol/L; SDS, 2%; EDTA, 5 mmol/L; aprotinin, 5 mg/ml; leupeptin, 1 mg/ml; pepstatin, 0.7 mg/ml; N-tosyl-L-lysine chloromethylketone, 50 mg/ml; N-tosyl-L-phenylalanine chloromethylketone, 100 mg/ml; soybean trypsin inhibitor, 100 mg/ml; phenylmethylsulfonyl fluoride, 1 mmol/L) and then sonicated. The cellular debris were pelleted by two 20-minute centrifugations (15,000 x g) at 4°C. The protein concentration was determined by the BCA protein assay (Pierce, Rockford, IL) and aliquots were stored at -80°C. Equal amounts of proteins were heated at 95°C for 4 minutes in SDS, dithiothreitol, and ß-mercaptoethanol containing sample buffer and fractionated by 12% SDS-PAGE. After transfer onto a nitrocellulose membrane, FGF-2 was detected by using a rabbit anti-FGF-2 polyclonal antibody (Santa Cruz Biotechnology) at a dilution of 1:500. Immune complexes were revealed by means of a horseradish peroxidase-conjugated anti-rabbit IgG antibody and an enhanced chemiluminescence kit (Amersham).
Hypoxia Assay
Pimonidazole binding has been used to assess changes in hepatic tissue oxygenation.15 Nitroimidazole compounds such as pimonidazole are reductively activated and covalently bound to macromolecules in cells at low oxygen concentration.16 In brief, the rats were injected with pimonidazole (120 mg/ml intravenously) 1 hour before killing at 1 week, 2 weeks, and 4 weeks following bile duct ligation or sham operation. Liver tissue was formalin-fixed and embedded in paraffin. Pimonidazole adducts were detected in formalin-fixed paraffin-embedded tissues with a biotin-streptavidin-peroxidase indirect immunostaining method modified for rat liver as previously described, using a monoclonal antibody provided by J. A. Raleigh.16
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Bile duct ligation triggered major structural changes in the liver
architecture, as previously described.3,17
After 3 days,
ductular reaction was readily detectable, and after 3 weeks, the
ductular reaction was intense and associated with extensive fibrosis.
After 7 weeks, biliary cirrhosis was present. Increased cell
proliferation in response to cholestatic liver injury occurred in both
parenchymal and nonparenchymal liver cells, as established before. The
kinetics of cell proliferation as ascertained by Ki67 immunolabeling
varied according to cell types: bile duct epithelial cells were the
first to proliferate and reached a plateau at 3 days, whereas the
proliferation of hepatocytes peaked at 1 week. The rate of positive
nuclei in bile ducts raised from 1.5% in normal liver to 38% at 3
days, while that of hepatocytes raised from 1.5% in normal liver to
18% at 1 week. Proliferation of the vascular endothelial cells lining
periportal vessels was determined on the same liver sections by the
number of Ki67-positive nuclei in these cells. While no significant
change was noted up to 1 week, inasmuch as 1.79 ± 1.79%,
6.25 ± 2.19% (ns) and 13.01 ± 2.98% (ns) of nuclei were
positive in normal rats, 3 days and 1 week after bile duct ligation,
respectively, an increase in the rate of Ki67-labeled nuclei occurred
thereafter and reached 27.86 ± 3.59% at 2 weeks
(P < 0.05 as compared with normal, 3 days and 1
week) (Figure 1, A and B)
.
|
and central veins (not shown). One week after BDL, there was no
significant change in vascular density (Figure 2B)
. In contrast, after
3 weeks and 6 weeks, there was a marked increase in vWF-labeled vessels
that were located within ductular reaction (Figure 2, C and D)
. The
results are entirely consistent with the time course of vascular
endothelial cell proliferation as shown in Figure 1
. Similarly, the
proliferation of bile duct epithelial cells was followed by a
significant increase in bile duct sections at 1 week. The
quantification of vascular density is shown in Figure 2E
(left) in
parallel with that of ductular proliferation (Figure 2E
, right), that
was assessed by the same method. However, both proliferation events
occurred at different time points.
|
On liver sections from normal rats, VEGF and FGF-2 immunolabeling
was detected in the first row of perivenular hepatocytes (Figure 3, A and B)
. In addition, 3% of
hepatocytes unevenly distributed within the lobule displayed FGF-2
reactivity. The labeling intensity was high in 1%, and moderate in 2%
of hepatocytes (Figure 3B)
. No expression was found in the other cell
types. Two weeks (data not shown) and 3 weeks (Figure 3C)
after BDL,
more than 95% of hepatocytes exhibited VEGF immunolabeling. Seven
weeks after BDL, an intense and homogeneous signal for VEGF was
observed in 100% of hepatocytes (Figure 3E)
. VEGF immunolabeling was
restricted to hepatocytes and was undetectable in nonparenchymal cells,
including bile duct epithelial cells and myofibroblasts. A similar
pattern of VEGF expression was obtained with the two different
polyclonal anti-VEGF antibodies that we have used (data not shown).
FGF-2 expression also increased after BDL but the time course and
distribution pattern differed from those of VEGF (Figure 3, D and F)
. A
significant signal for FGF-2 was detected in 49% (intense in 9%) of
hepatocytes 3 weeks after BDL (Figure 3D)
and in more than 95%
(intense in 35%) of hepatocytes after 7 weeks (Figure 3F)
.
|
. In contrast, the bands at 434 and 565 bp, corresponding to the
120- and 164-amino-acid isoforms respectively, increased by 1.5 to
2-fold 4 weeks after BDL. Moreover, the transcript of 164-amino acid
isoform became predominant within 2 weeks after BDL (Figure 4, A and B)
.
|
. Three isoforms of 18, 21, and 22 kd
were detected (Figure 5)
. The former corresponds to an AUG codon
initiated form while the 21- and 22-kd peptides are initiated at codons
CUG3 and CUG2, respectively.21
The 22-kd FGF-2 was the
main isoform expressed in all samples and the 31-kd FGF-2 initiated at
CUG1 codon22
was undetectable even after long time
exposures.
|
Induction of Liver Hypoxia
The demonstration of an intense hepatocellular overexpression of
VEGF following BDL prompted us to investigate the role of liver hypoxia
as a triggering event. In normal rats (Figure 6A
, top) pimonidazole adducts were
undetectable in the liver. One week after BDL, a patchy staining for
pimonidazole adducts was observed (data not shown). Two weeks and 4
weeks after surgery (Figure 6, B and C
, respectively; top),
pimonidazole adducts were detected in more than 95% of hepatocytes. In
normal rats (Figure 6A
, bottom) and at 1 week (data not shown), VEGF
was detected in perivenular hepatocytes. Pimonidazole adducts and VEGF
expression displayed the same pattern on adjacent liver sections 2
weeks (Figure 6B
, bottom) as well as 4 weeks after BDL (Figure 6C
,
bottom). Determination of PO2 in the peripheral arterial
blood from the same animals revealed no significant change in the
oxygen tension of rats 3 weeks, 4 weeks as well as 6 weeks after BDL,
as compared with normal rats.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It has been established that sinusoidal perfusion is impaired in cirrhosis, whatever its cause12,13 and regenerative nodules in human and experimental cirrhosis are constantly surrounded by a dense perinodular vascular plexus.1-3 These numerous and tortuous microvessels originate from intrahepatic vascular branches, progress together with the fibrous repair process and bypass the obstructed normal route.1 These observations suggest that an impaired oxygen delivery to the hepatocytes may occur in cirrhosis as a result of intrahepatic shunts or capillarization of the sinusoids, leading to hepatocyte hypoxia.23 In agreement with this view, we identified angiogenesis and hypoxia in cirrhotic tissues. However, our data indicate that hypoxia, VEGF induction, and angiogenesis precede the onset of cirrhotic lesions. Early hypoxia is compatible with the occurrence of liver blood supply impairment long before cirrhosis13.
The immunochemistry data showed an early induction of VEGF in hepatocytes. Furthermore, this VEGF expression paralleled an increase in the level of transcripts coding for the secreted isoforms VEGF120 and VEGF164 which act directly on the proliferation and migration of vascular endothelial cells. These isoforms also increase the permeability of microvessels to circulating macromolecules.19 In contrast to VEGF121 and VEGF165, human VEGF189 binds to the extracellular matrix and requires hydrolysis by proteases like plasmin to be activated.24 Of note, plasma urokinase-type plasminogen activator increases in liver cirrhosis25 and myofibroblastic hepatic stellate cells synthesize components of the plasminogen-activating system, generating plasmin that plays a key role in matrix remodeling.26 In addition, VEGF has other biological activities (such as induction of the expression of different proteases by endothelial cells27,28 and stimulation of endothelial cells and monocyte procoagulant activity29 ) that might also indirectly induce microvascular remodeling in the liver by promoting monocyte migration and adhesion of activated neutrophil.13,29 The time course of vessel proliferation after BDL was entirely consistent with an angiogenic response to VEGF induction.
By immunochemistry, we also found an up-regulation of FGF-2 was delayed by comparison with VEGF expression after BDL. Western blot analysis indicated that the 22-kd FGF-2 was the main isoform expressed in this experimental model. FGF-2 acts as a potent mitogen and an inducer of cell migration in different cell types, including endothelial cells and fibroblasts.30 Low and high molecular weight isoforms of human FGF-2 generated by alternative initiation of translation have been described.21,31 It has been suggested that the low molecular weight form (18 kd) might modulate cell motility and proliferation through interaction with its cell surface receptor while the high molecular weight isoforms (22 to 24 kd) might only act as a mitogen through an intracellular mechanism.20 FGF-2 expression has been shown to be associated with fibrous septa in both human and experimental chronic liver diseases.32 In addition, a synergism between FGF-2 and VEGF in the induction of angiogenesis and activated hepatic stellate cell proliferation has been demonstrated33. Finally, FGF-2 has been shown to inhibit endothelial cell apoptosis.34 Altogether, these data suggest that FGF-2 might contribute to maintain a vascular proliferative reaction previously induced by VEGF.
Our results strongly argue for the role of hepatocyte hypoxia as an early triggering event in the induction of VEGF in this experimental model. It has been previously demonstrated that hypoxia induces VEGF at both mRNA and protein level.35,36 Furthermore, hypoxia-induced up-regulation of VEGF expression has been shown to involve both a transcriptional activation37 and a stabilization of transcripts.38 However, we cannot definitively exclude an additional role of factors produced by activated inflammatory cells,39,40 since certain cytokines or growth factors (ie, EGF, HGF, PDGF, TGF-ß) are able to stimulate VEGF expression in specific cell types.41-44 Nevertheless, infiltration by inflammatory cells remained moderate in our experiments. In addition, the induction of VEGF in hepatocytes was associated with the concomitant occurrence of hypoxia in the same areas. In contrast, the expression of FGF-2, which has been shown to be insensitive to hypoxia,43,45 was induced later and did not respond to the same distribution pattern suggesting that the inducing factors involved in the expression of the two angiogenic factors are different.
In conclusion, our results demonstrate for the first time the sequential induction of two major angiogenic factors, VEGF and FGF-2, during biliary-type liver fibrogenesis. They suggest that hypoxia might be a major factor implicated in the induction of VEGF and in the marked angiogenesis occurring at an early stage before the onset of cirrhotic lesions. Further studies are now required to determine the importance of these factors in liver diseases and whether their modulation might influence the progression of liver tissue repair.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by grants from the Delegation à la Recherche Clinique of Assistance Publique-Hôpitaux de Paris (project CRC96181) and U.S. National Cancer Institute Grant R42 CA68826.
Accepted for publication June 3, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L. Fouassier, M. Beaussier, E. Schiffer, C. Rey, V. Barbu, M. Mergey, D. Wendum, P. Callard, J.-Y. Scoazec, E. Lasnier, et al. Hypoxia-induced changes in the expression of rat hepatobiliary transporter genes Am J Physiol Gastrointest Liver Physiol, July 1, 2007; 293(1): G25 - G35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Novo, S. Cannito, E. Zamara, L. V. di Bonzo, A. Caligiuri, C. Cravanzola, A. Compagnone, S. Colombatto, F. Marra, M. Pinzani, et al. Proangiogenic Cytokines as Hypoxia-Dependent Factors Stimulating Migration of Human Hepatic Stellate Cells Am. J. Pathol., June 1, 2007; 170(6): 1942 - 1953. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kanematsu, S. Osada, N. Amaoka, S. Goshima, H. Kondo, H. Nishibori, H. Kato, M. Matsuo, R. Yokoyama, H. Hoshi, et al. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma and the Surrounding Liver: Correlation with Angiographically Assisted CT Am. J. Roentgenol., December 1, 2004; 183(6): 1585 - 1593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. L. Ward, A. L. Haninec, P. Van Slyke, J. G. Sled, C. Sturk, R. M. Henkelman, I. R. Wanless, and D. J. Dumont Angiopoietin-1 Causes Reversible Degradation of the Portal Microcirculation in Mice: Implications for Treatment of Liver Disease Am. J. Pathol., September 1, 2004; 165(3): 889 - 899. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. V. Smith, A. Nyska, and C. Portier Application of a Statistical Dynamic Model Investigating the Short-Term Cellular Kinetics Induced by Riddelliine, a Hepatic Endothelial Carcinogen Toxicol. Sci., August 1, 2004; 80(2): 258 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H Yoshiji, S Kuriyama, J Yoshii, Y Ikenaka, R Noguchi, D J Hicklin, Y Wu, K Yanase, T Namisaki, M Yamazaki, et al. Vascular endothelial growth factor and receptor interaction is a prerequisite for murine hepatic fibrogenesis Gut, September 1, 2003; 52(9): 1347 - 1354. [Abstract] [Full Text]
|
|
|
|
|
|
T. V. Masyuk, E. L. Ritman, and N. F. LaRusso Hepatic Artery and Portal Vein Remodeling in Rat Liver: Vascular Response to Selective Cholangiocyte Proliferation Am. J. Pathol., April 1, 2003; 162(4): 1175 - 1182. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Ohashi, A. Shimizu, Y. Masuda, H. Kitamura, M. Ishizaki, Y. Sugisaki, and N. Yamanaka Peritubular Capillary Regression during the Progression of Experimental Obstructive Nephropathy J. Am. Soc. Nephrol., July 1, 2002; 13(7): 1795 - 1805. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. M. Mazure, C. Chauvet, B. Bois-Joyeux, M.-A. Bernard, H. Nacer-Cherif, and J.-L. Danan Repression of {alpha}-Fetoprotein Gene Expression under Hypoxic Conditions in Human Hepatoma Cells: Characterization of a Negative Hypoxia Response Element That Mediates Opposite Effects of Hypoxia Inducible Factor-1 and c-Myc Cancer Res., February 1, 2002; 62(4): 1158 - 1165. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Corpechot, V. Barbu, D. Wendum, N. Chignard, C. Housset, R. Poupon, and O. Rosmorduc Hepatocyte Growth Factor and c-Met Inhibition by Hepatic Cell Hypoxia : A Potential Mechanism for Liver Regeneration Failure in Experimental Cirrhosis Am. J. Pathol., February 1, 2002; 160(2): 613 - 620. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Pillebout, M. Burtin, H. T. Yuan, P. Briand, A. S. Woolf, G. Friedlander, and F. Terzi Proliferation and Remodeling of the Peritubular Microcirculation after Nephron Reduction : Association with the Progression of Renal Lesions Am. J. Pathol., August 1, 2001; 159(2): 547 - 560. [Abstract] [Full Text]
|
|
|
|
|
|
C. C. Wykoff, N. J. P. Beasley, P. H. Watson, K. J. Turner, J. Pastorek, A. Sibtain, G. D. Wilson, H. Turley, K. L. Talks, P. H. Maxwell, et al. Hypoxia-inducible Expression of Tumor-associated Carbonic Anhydrases Cancer Res., December 1, 2000; 60(24): 7075 - 7083. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
